SemaWal Semantic network resolver

SemaWal is a semantic network resolver in python

Developper: Ali AOUF

Description

SemaWal Is a semantic network resolver developed as a python library. It allows the creation of networks through manual coding or a CSV file. It supports many types of connections between nodes. In addition to extracting knowledge (mainly relations) between two or more nodes in a network, it can find paths between nodes in a given network.

Features:

• Support for multi-relations(connections between nodes).
• Support for three types of relations: Strict, regular, extend(inheritance).
• Support for positive and negative relations.
• Support for CSV files.
• Support for Sub-networks.
• Node functions:
  • Fetch all the related nodes and all the relations that a given node has.
  • Fetch nodes that are related to a given node.
  • Fetch relations that relate two given nodes has.
  • Fetch all the relations that a network contains.
  • Fetch all the relations you can get starting from a given node until reaching a given depth.
  • Check if two given nodes are directly connected by a given relation.
  • Check if two given nodes are connected through a path of nodes.
  • Check if two given nodes are connected through a path of nodes by a given relation.
  • Draw a path between two given nodes.

In progress features

• A GUI interface
• Nodes properties
• An option to ignore certain relations

Usage

install

pip install semawal

[requirement]

import

>>> from semawal.node import Node
>>> from semawal.net import Net
>>> from semawal.csv_parser import parser

Examples

Detailed examples and features in test.py

• Creating nodes

  >>> nodeA = Node(name="A")
  [!] Created node A
  >>> nodeB = Node(name="B")
  [!] Created node B
  >>> nodeC = Node(name="C")
  [!] Created node C
  

• Linking two nodes

  >>> ## One-way relation: Ex: A has B
  >>> # mode (int): 0=negative, 1=positive
  >>> # power (int): 0=none strict, 1=strict
  >>> nodeA.link(attribute="has", node=NodeB, mode=1, power=1)
  >>> ## Mutual Link: Ex: A is B and B is A
  >>> nodeA.mutual_link(attribute="is", node=NodeB, mode=1, power=1)
  >>> ## Inheritance (A inherits B relations): Ex: A extends B
  >>> nodeA.extends(node=NodeC)
  

• Node functions

  >>> ## Return all the nodes that are connected to a given node.
  >>> # all (boolean): consider negative relation(relations with mode=0) like "is not", "has not"... 
  >>> nodeA.connections(all=true)
  {<node.Node object at 0x7ff0d8d32310>, <node.Node object at 0x7ff0d8cfcfd0>}
  >>> ## Return relations with a given node
  >>> nodeA.relationsWith(node=nodeB)
  ['is']
  >>> ## Print all the relations that a node has
  >>> nodeA.showLinks()
  A       |   is  |   B  mode:  1  strict:  1
  A       |   extends     |   C  mode:  1  strict:  1
  >>> ## Check if there is a direct relation between two nodes
  >>> nodeA.check(attribute="is", node=nodeB, mode=1)
  True
  

• Creating a network

  >>> myNet = Net(name="MyNet")
  [!] Created network MyNet
  

• Importing a network from a csv file

  >>> n = parser.read("link_test.csv")
  [!] Created network  link_test
  

• Adding a node to a network

  >>> myNet.add(nodeA, nodeB, nodeC)
  

• Get a node from a network

  >>> myNet.getNode("A")
  

• Get a list of all the nodes in a network

  >>> myNet.getNodeskeys()
  ['A', 'B', 'C']
  

• Fetching all the relations in a network

  >>> myNet.randomSearch()
  A       |   is  |   B  mode:  1  strict:  1
  A       |   extends     |   C  mode:  1  strict:  1
  B       |   is  |   A  mode:  1  strict:  1
  

• Fetching all the relations in a starting from a giving node until reaching a given depth

  >>> myNet.search(node=nodeB, depth=3)
  B       |   is  |   A  mode:  1  strict:  1
  A       |   is  |   B  mode:  1  strict:  1
  A       |   extends     |   C  mode:  1  strict:  1
  

• Get the path that relats two nodes

  >>> myNet.areConnected(nodeA=nodeA, nodeB=nodeB)
  B       |   is  |   A  mode:  1  strict:  1
  A       |   is  |   B  mode:  1  strict:  1
  [<node.Node object at 0x7f0571bbb4c0>, <node.Node object at 0x7f0571ac2310>]
  >>> # that was [nodeA, nodeB]
  

• Printing the path that relats two nodes

  >>> myNet.drawPath(nodeA=nodeA, nodeB=nodeB)
  Printing the path:
  A       |   ['is']      |   B